PORTLAND, Ore. -- An expanding suite of algorithms for fusing the outputs from Xtrinsic family of MEMS chips is designed to enable designers to develop sensor fusion algorithms from Freescale Semiconductor application notes or to license a sensor-hub software solution from motion algorithm specialists Movea Inc.
"Sensor fusion covers a variety of techniques ranging from trading off the strengths of one sensor to compensate for weakness in another sensor [to] improving the quality and noise in computed results by using known data redundancies among sensors," said Mike Stanley, IC architect and systems engineer at Freescale (Austin, Texas).
Stanley writes a sensor blog on the latest sensor fusion algorithms and Freescale applications notes for implementing them. One technique enables sensors to minimize power consumption by turning off power to the application processor during periods of inactivity, then restoring power when motion is detected by a MEMS sensor.
Simple applications like compensating for tilt when using magnetometer can be relatively straightforward for engineers to program, but increasingly complex sensor fusion algorithms are a tougher problem when using them to analyze, auto-calibrate and cross-compensate for errors in raw data outputs from multiple MEMS sensors. These sensors include accelerometers, gyroscopes, magnetometers, altimeters along with temperature- and proximity-sensors.
Freescale if offering an expanded sensor toolbox for a fee, and engineers can also license a sensor hub solution from Movea that off-loads the application processor by running core algorithms on a separate Freescale ColdFire microcontroller.
Movea is co-marketing its MotionCore sensor fusion platform with Freescale for use with its Coldfire microcontrollers, said Tim Kelliher, director of customer solution architectures at Movea (Pleasanton, Calif.) "By incorporating sensor fusion into our separate hub, instead of running sensor fusion algorithms on the application processor, designers can create motion-aware applications that are agnostic to hardware," Kelliher added in a statement.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.